Image-Based Reverse Engineering and Visual Prototyping of Woven Cloth

Abstract

Realistic visualization of cloth has many applications in computer graphics. An ongoing research problem is how to best represent and capture cloth models, specifically when considering computer aided design of cloth. Previous methods produce highly realistic images, however, they are either difficult to edit or require the measurement of large databases to capture all variations of a cloth sample. We propose a pipeline to reverse engineer cloth and estimate a parametrized cloth model from a single image. We introduce a geometric yarn model, integrating state-of-the-art textile research. We present an automatic analysis approach to estimate yarn paths, yarn widths, their variation and a weave pattern. Several examples demonstrate that we are able to model the appearance of the original cloth sample. Properties derived from the input image give a physically plausible basis that is fully editable using a few intuitive parameters.

Fitting procedural yarn models for realistic cloth rendering

“…Although these models provide high-quality renderings for fabrics viewed from a distance, they lack the power to closely reproduce the appearance of fabrics with thick yarns or fuzzy silhouettes, or the generality to handle different fabrics with wildly varying appearances. Recent advances in cloth appearance modeling have led to the development of volumetric [Xu et al 2001;Jakob et al 2010;Zhao et al 2011;Zhao et al 2012;Zhao et al 2013] and fiber-based [Schröder et al 2015;Khungurn et al 2015] cloth models. Unlike traditional methods, they describe fabrics at 3D volumes with varying densities (volumetric) or collections of fiber curves (fiber-based).…”

“…Compared to traditional surface-based methods, these models are better at capturing the thickness and fuzziness of fabrics, which add significantly to their visual realism, and thus have brought the quality of computed rendered cloth to the next level. However, these techniques either provide automated creation of highly realistic models (achieved by measuring real-world samples) but carry little highlevel synthetic information (e.g., yarn twisting) [Zhao et al 2011;Khungurn et al 2015], or offer compact procedural representations but rely on manually configured models which require nontrivial parameter tweaking to match reality closely [Schröder et al 2015].…”

“…SIGGRAPH '16 Technical Paper, July 24 -28, 2016, Anaheim, CA, ISBN: 978-1-4503-4279-7/16/07 DOI: http://dx.doi.org/10.1145 Representing the richness of fabrics in virtual models has actively driven research in a variety of fabric appearance models. Recently, micro-appearance based fabric models [Zhao et al 2011;Zhao et al 2012;Khungurn et al 2015;Schröder et al 2015] that explicitly capture a fabric's micro-scale geometry have been introduced. Compared to traditional methods, these models describe fabrics at unprecedented detail, and thus offer superior generality as the complex appearance of a fabric is largely a direct consequence of its small-scale structures.…”

“…Recently, Schröder et al [2015] introduced a procedural yarn model to computer graphics leveraging state-of-the-art results from this field [Morris et al 1999;Tao 1996;Keefe 1994 which can be edited easily. However, how to set parameter values for this model to closely match real fabrics remains a significant challenge.…”

“…However, how to set parameter values for this model to closely match real fabrics remains a significant challenge. In [Schröder et al 2015], creating procedural yarns was a manual process. But this fitting can be tricky.…”